Am. J. Hum. Genet. 59:1074-1083, 1996

Myosin VIIA Mutation Screening in 189 Usher SyndromeType 1 PatientsM. D. Weston,1 P. M. Kelley,' L. D. Overbeck,' M. Wagenaar,3 D. J. Orten,' T. Hasson,4'5Z.-Y. Chen,6'7 D. Corey,6'7 M. Mooseker,4'5'6 J. Sume i, C. Cremers,3 C. M Iller,8S. G. Jacobson,9 M. B. Gorin,'° and W. J. Kimberling

'Department of Genetics, Boys Town National Research Hospital, and 2Department of Microbiology and Pathology, University of NebraskaMedical Center, Omaha; 3Department of Otorhinolarynogology, University Hospital Nijmegen, Nijmegen; Departments of 4Biology,5Pathology, and bCell Biology, Yale University, New Haven; Howard Hughes Medical Institute, Massachusetts General Hospital andProgram in Neuroscience, Harvard Medical School, Boston; 8Department of Otolaryngology, University of Linkoping, Linkoping, Sweden;9Scheie Eye Institute, University of Pennsylvania, Philadelphia; and "0Departments of Ophthalmology and Human Genetics, University ofPittsburgh, Pittsburgh

SummaryUsher syndrome type lb (USHiB) is an autosomal reces-sive disorder characterized by congenital profound hear-ing loss, vestibular abnormalities, and retinitis pig-mentosa. The disorder has recently been shown to becaused by mutations in the myosin Vha gene (MYO7A)located on llql4. In the current study, a panel of 189genetically independent Usher I cases were screened forthe presence of mutations in the N-terminal coding por-tion of the motor domain of MYO7A by heteroduplexanalysis of 14 exons. Twenty-three mutations werefound segregating with the disease in 20 families. Of the23 mutations, 13 were unique, and 2 of the 13 uniquemutations (Arg2l2His and Arg2l2Cys) accounted forthe greatest percentage of observed mutant alleles (8/23, 31%). Six of the 13 mutations caused prematurestop codons, 6 caused changes in the amino acid se-quence of the myosin VI1a protein, and 1 resulted ina splicing defect. Three patients were homozygotes orcompound heterozygotes for mutant alleles; these threecases were Tyr333Stop/Tyr333Stop, Arg212His-Arg3O2His/Arg212His-Arg3O2His, and IVS13nt-8c-/Glu450Gln. All the other USHlB mutations observedwere simple heterozygotes, and it is presumed that themutation on the other allele is present in the unscreenedregions of the gene. None of the mutations reported herewere observed in 96 unrelated control samples, althoughseveral polymorphisms were detected. These results addthree patients to single case reported previously wheremutations have been found in both alleles and raises thetotal number of unique mutations in MYO7A to 16.

Received June 4, 1996; accepted for publication August 21, 1996.Address for correspondence and reprints: Dr. William J. Kimberling,

Center for Hereditary Communication Disorders, Boys Town Na-tional Research Hospital, 555 North 30th Street, Omaha, NE 68131.E-mail: kimber@boystown.org© 1996 by The American Society of Human Genetics. All rights reserved.0002-9297/96/5905-0016$02.00

Introduction

Phenotypically, there are three types of Usher syndrome:I, II, and III. Type I has a profound congenital hearingloss, vestibular areflexia, and retinitis pigmentosa (RP).Type II has a mild hearing loss in the low frequencies,sloping to severe in the high frequencies, normal vestibu-lar responses, and RP. Type III has a progressive loss,variable vestibular problems, and RP. At least three dif-ferent genes are responsible for the more severe type IUsher syndrome. These genes are located on 1 1q(USH1B [Kimberling et al. 1992]), 11p (USH1C [Smithet al. 1992]), and 14q (USHlA [Kaplan et al. 1992]).There are no known clinical differences between thesethree subtypes of Usher type I, and they can be distin-guished only by linkage in informative families. Usherlb is the most common subtype, accounting for 70%-80% of all type I cases. Usher Ia accounts for mostof the other cases, while Usher Ic has been observedexclusively in people with French Acadian ancestry.The Usher syndrome type lb (USH1B) subtype has

been localized to 1ql4 (Kimberling et al. 1992). Re-cently, a 3.5-kb partial cDNA from human testis codingfor myosin VIIa (MYO7A) has been cloned from thisregion (Hasson et al. 1995). Deleterious mutationswithin the coding sequence are associated with familiessegregating the disorder (Weil et al. 1995). Similarly,mutations in the homologous mouse myo7 gene causethe shaker-1 (sh-i), whose phenotype is deafness with-out retinal dystrophy (Gibson et al. 1995). DFNB2 isa human neurosensorial recessive deafness gene that isunassociated with retinal degeneration but maps to thesame location on llq as USH1B (Guilford et al. 1994),raising the possibility that some mutations in MYO7Amay cause a human phenotypic equivalent of sh-1.We undertook a mutation screening study of type I

Usher cases by heteroduplex analysis (HA) of 14 exonsto gain insights into the position and nature of differentMYO7A mutations and to lay a foundation for studies

1074

Weston et al.: Mutation Analysis of Myosin VIIa

into whether different mutations result in different phe-notypes.Usher syndrome is responsible for -5% of all congen-

ital deafness (Kimberling and Moller 1995), and amethod of rapid and inexpensive screening of youngpatients with congenital profound deafness would bebeneficial. Early identification of Usher cases will affecta child's education as well as enhance his/her safety byalerting parents about potential problems with night andperipheral vision. In order to evaluate different strategiesfor such large-scale mutation screening, the frequencyof specific myosin-VIIa mutations and their distributionrelative to different populations and ethnic groups mustbe determined.

This is a report on 13 new MYO7A mutations thatwere observed exclusively in a panel of Usher Tb patients.The nature of each mutation and the predicted mecha-nisms for MYO7A inactivation is discussed.

Subjects and Methods

SubjectsThe cases studied here were collected as part of a

long-standing study on the genetics of Usher syndrome.Patients were considered affected with Usher type I, onthe following basis: congenital severe to profound hear-ing loss, vestibular areflexia, and RP (confirmed in atleast one affected family member by an electroretino-gram). The cases studied were generally of two types;informative families compatible with linkage to 11q14markers, which are presumed to have Usher Ib, or single-ton cases with a phenotype fitting the diagnostic criteriastated above. A linkage diagnosis was not possible forthe singleton cases; therefore, we are not absolutely cer-tain that they are Usher type Tb. However, data on thefrequency of linked families indicates that >80% of allpatients with type I Usher syndrome are linked to 11qand are presumed to carry MY07A mutations. Usher Ifamilies failing to link to 11q markers flanking theUSH1B locus were not studied.

7D11 DNAA cosmid array produced from the partial Sau3AI

digestion of CEPH YAC 802a5 (P. M. Kelley, unpub-lished results) was screened using a 4.6-kb mouse cDNA,ET58 (Gibson et al. 1995). Cosmid 7D1 1 was identifiedand a 4.2-kb and a 6.0-kb EcoRI fragments hybridizingto ET58 were subcloned. The sequence was obtained bygenerating deletion clones using the Erase-a-Base Sys-tem® (Promega) and primer walking using oligonucleo-tides synthesized on a Cruachem PS250. These subclonescontained the 10 exons from the MYO7A head regionreported by Weil et al. (1995).

Intron/Exon boundaries of 4 additional 5' exons wereidentified by sequencing 7D1 1 template using primers

designed from the sequence of MYO7A cDNA (Hassonet al. 1995). These 14 exons correspond to the first1,957 bases of the cDNA (Accession number U34227),and the sequence information was used to design theprimers for mutation screening.

Mutation DetectionIntronic PCR amplification primers flanking each

exon (table 1) were used for screening mutations by HA(Keen et al. 1991). The exceptions were exons 4 and 9.For each of these exons, one primer spanned an intron/exon boundary. PCR reaction mixtures contained 200ng of genomic DNA, 100 pmol of each exon-specificprimer pair, 1.5 mM MgCl2, and dNTPs (200 ,uM) ina total volume of 100 pl. Amplification was carried outusing a hot-start method to minimize false priming: 96°Cfor 5 min, followed by a soak at 80°C, then 2.5 U ofTaq DNA polymerase (Perkin Elmer) and dNTPs wereadded followed by 37 cycles of 30 s at 95°C, 30 s at60°C, and 45 s at 72°C, with a final incubation for 5min at 72°C.Homozygous and heterozygous mutations were de-

tected as follows: PCR products amplified from genomicUSH1B samples with 7D1 1 cosmid were mixed, heatedat 95°C for 3 min, and cooled to 37°C over a 20-30-min period. The reannealed reaction products were elec-trophoresed through 35-cm x 43-cm X 1-mm 1 xMDETM gels (FMC) at 850 V for 16-24 h. PCR reac-tions were repeated on the heteroduplex positive sam-ples and all available family members. HA was doneboth with and without 7D1 1 PCR control, to determinewhether the patient was homozygous or heterozygous.Genomic DNA from a set of 96 normal genetically inde-pendent control samples were also analyzed to help dif-ferentiate pathologic mutations from probable polymor-phisms.

Amplified exons from patients were purified by theWizardTM PCR clean-up kit (Promega) and sequencedusing the fmol® sequencing kit (Promega) using appro-priate end-labeled primers. If direct sequencing did notunequivocally identify the DNA sequence change, thePCR product was cloned into a plasmid vector (AT clon-ing kit, Invitrogen) to separate individual alleles priorto sequencing.

Results

Of 378 USH1B chromosomes surveyed for mutationsby HA, 13 different mutations were found in a total of23 USH1B chromosomes from 20 probands. Table 2summarizes the mutations observed. Codon numberingstarts with the first in-frame methionine of MYO7A(Hasson et al. 1995; Weil et al. 1996; Chen et al., inpress). Figure 1 shows the relative locations of each mu-tation along a linear representation of the protein.

1075

Am. J. Hum. Genet. 59:1074-1083, 1996

Table 1

MY07A PCR Primers

PCR Fragment SizeExon and Primer Sequence (bp)

GAG*CAG*CCT*CTG-GAC-ATT AGT-CCCA TGA-CTC CTC*CAC*ATC*CA

CAG*AAG*CAT*GAC-ATG-GTC TCT*CTGCA-GGA* ATF7 TFTC CAA*GAG-AAC AC

CCT TGT-GTG-GTT*GGC TCA-GATCG*ATC-ATC TCC CTG CAT* CA

ACC-ATC*CTG-ACT-GAA-CCC*TGTTCA*CGT*AGA-TGA-GGT-GGT*CC

AGT*CAG*AGT-CTC-CCA*CAT-GGAATC*AGC*AGG AGC GCA*CAG-T

TGT*GGG TTG-TGA-CAG- GTC CTTGC*GGA*GTG AGT AGG GTG TC

ACC AGA-GTT-CCG*AGG*GTG-AGG* GGC* CTG- GGT* CTA TTC-

CAC*TGT*GCC-CAC-ATT*TTC-AGGCC*AGC TCC*CTA CAA*TCC-TA

CCA-CCA*GGG TAA*CTG*CAT-AACAAG*AGG*GCA-CGT GCT*AGG*A

CCT GGG* GAA* GCA-TT*AGT GCATGA'ACA-GAA*CTC TCC-AAA-CCA*GA

GGG CTG TCA AGG*AGA*GAG-AAGCCCGATC*AAG AGC-CCA-ACT*TC

GGT TTC ACA CGG GCAC-TlT- GGCC AGG-AGC-CAA* CTA* AAT-GTG

GGA* GGG AGG TGG* ACT* TGA* CAAA*GCA* GGG-AAG* GAA* GCT-GT

CCA GAG AGC GCC TAT-GTG AGCTFT*ATT* CCT' GGG* CTG-GAG-CA

441

206

312

328

310

241

260

195

272

354

393

365

432

349

Missense MutationsSix different missense mutations were found in 13

Usher I patients. Seven Usher lb patients were ob-served with missense mutations in exon 7. These are

the Arg212His and Arg2l2Cys mutations originallyobserved in two Usher lb cases as heterozygotes (Weilet al. 1995). They are believed to be pathologic because(1) they occur within an evolutionarily conserved hep-tapeptide sequence that is invariant in all myosins stud-ied and (2) neither has been observed in a series ofcontrol persons. Figure 2 shows the heteroduplex pat-

tern observed in two Usher Ib families with theArg212His mutation. PCR sequencing of exon 7 was

carried out on at least one heteroduplex positive pa-tient in each family to verify that the mutation was dueto either a G--A(CGT--*CAT) or a C-'T(CGT-+TGT)transition in codon 212. Furthermore, an AluI site is

created by the Arg212Cys (C-'T) mutation. Amplifiedexon 7 fragments were digested with AluI, and allArg212Cys (C-1T) mutations showed the predictedrestriction fragment sizes following agarose gel elec-trophoresis.

1:MY07-lUMY07-lL

2:MY07-2UMY07-2L

3:MY07-3UMY07-3L

4:MY07-4UMY07-4L

5:MY07-SUMY07-5L

6:MY07-6UMY07-6L

7:MY07-7UMY07-7L

8:MY07-8UMY07-8L

9:MY07-9UMY07-9L

10:MY07-IOUMY07-1OL

11:MY07-11UMY07-1 1L

12:MY07-12UMY07-12L

13:MY07-13UMY07-13L

14:MY07-14UMY07-14L

1076

Weston et al.: Mutation Analysis of Myosin VIla

Table 2

Summary of Mutations in 189 Usher-Type IB Cases for MY07A Exons 1-14

Patient Status and Codon No. ofChange Exon Mutation Chromosomes Origin

Heterozygotes:Cys31Stop 3 TGC-*TGA 1 Sweden

1 Holland75delG(FS) 4 del GAC 1 Australia120delC(FS) 5 del CGCCAG 1 HollandArg212His 7 CGT-.CAT 1 USAArg212Cys 7 CGT-*TGT 2 USA

1 Holland1 Belgiuma

Arg3O2His 9 CGC-CAC 1 Holland1 Sweden

Glu314Stop 9 GAG-+TAG 1 USA

Arg212His-Arg302His 7 and 9 See above 1 Holland468+Gln 13 insert GCA 2 USAProSO3Leu 13 CCC-)CTC 1 USA532delA(FS) 14 del CAC 1 USA

Homozygotes:Arg212His-Arg302His 7 and 9 See above 2 BelgiumaTyr333Stop 9 TAT-+TAG 2 USA

Compound Heterozygote:Glu450Gln/ 13 and GAG-+CAG 1 USAIVS13nt-8c-+g 14 ctccccag-tccccag 1

a These families were collected from the West-Flanders region, a Dutch-speaking part of Belgium.

Figure 2 also shows the heteroduplex pattern of aArg212His mutation (exon 7) in cis with an Arg3O2Hismutation (exon 9) in two Usher lb families. Affectedsiblings in a Dutch family (1515) are homozygous for adouble mutation at both codons, while the affected sib-lings in Flemish family (1062) show only paternal inheri-tance of both mutations. Arg3O2His and Arg212Hishave both been observed singly in affected persons. Nei-ther of the two mutations were observed in controlseither singly or as double mutations.

Three exon 13 mutations were observed. Figure 3Ashows the segregation of the heteroduplex patterncaused by a GCA insertion at codon 468 (468+Gln) infamily 985. This three-base insertion was also observedin one additional singleton case as a heterozygote. Theexon 13 mutation ProSO3Leu was observed in family734 (Fig. 3B). This is an unusual family because theproband also has Crouzon syndrome. His sister, whocarries the Pro5O3Leu mutation, was reported to havea progressive bilateral hearing loss but does not haveRP. The third exon 13 mutation occurred in trans witha splice-site mutation and is explained in the followingparagraph.Splice Site MutationA C--G transversion at position -8 in the intronic

splice-acceptor region preceding exon 14 was observed

in one family, 218 (fig. 3B). The predicted effect of sucha mutation is the disruption of the polypyrimidine tractcharacteristic of the majority of vertebrate 3' splice sites.Normal splicing would be inhibited by impaired bindingof various splicosome components that interact with thepolypyrimidine tract (Krawczak et al. 1992). The pro-band from this family also inherited a G--C transversionin exon 13 at codon 450 (Glu450Gln) from the otherparent (fig. 3B). The contribution of mutant alleles fromeach parent in this family is consistent with the hypothe-sis that the presence of both mutated alleles in the pro-band is responsible for the Usher type I phenotype.

Nonsense MutationsSix mutations that directly or indirectly, through a

frameshift, lead to premature stop codons were foundin seven Usher lb cases. Two unrelated cases (one fromSweden and the other from the United States) werefound to be heterozygous for a C-.A transversion inexon 3 at codon 31, causing a Cys-+Stop, (TGC-+TGA).The U.S. family has Scandinavian ancestry, and so onemust consider the possibility that the two are derivedfrom the same ancestral mutation.A deletion of a single G in a GGGGG string in exon

4 causes a frameshift starting a codon 75 and continuesfor 30 amino acids and ends at a UGA stop codon. This

1077

Am. J. Hum. Genet. 59:1074-1083, 1996

;;din

clx

D75f~sdeiGR120fsdeIC

R212H/ / |Rt212C / Y333X

R302H E314X

Motor Head motifs coil

aAL

11'

E4 0o0

H532fsdeIAIVS13-8c->g

PS03L466+Q

Figure 1 Partial domain structure of myosin VIla protein with relevant motifs, which define the myosin superfamily of proteins. The 3'tail domains specific to myosin VIla are omitted from this figure. The relative linear location of each of the mutations are shown. Boldfacedmutations are missense, and italicized mutations are frameshift, nonsense, and splice-site mutations. The horizontal black line inside the motor

head represents the region within the conserved motor domain covered by the 14 exons screened for mutations.

mutation was observed in one USH1B family with themother and her two affected sons being heterozygousfor the deletion.A single base deletion in exon 5 between codons 120

and 121 (CGCCAG-+CGCAG) causes a frameshift end-ing in an UGA stop after 24 amino acids. This mutationwas observed in the heterozygous state in two affectedsiblings and their mother.A single base deletion in exon 14 at codon 532 (CAC)

occurred in one Usher lb family. The mother and threeaffected sibs were all heterozygous for this deletion. Themutation causes a frameshift of 78 codons and ends at

a UGA within the actin binding domain of the myosinVIIa molecule.Two Usher 1 cases harbored separate nonsense muta-

tions in exon 9. The first was a heterozygous G-+T trans-

version in codon 314 (GAG--TAG) carried in an isolatedUSH1B case. The other nonsense mutation is a homozy-gous T-+G transversion in codon 333 (fig. 4).

PolymorphismsSeveral base substitution polymorphisms were de-

tected by HA and sequencing in both affected andcontrol sample groups (table 3). Weil et al. (1995)reported polymorphisms in exons 5 and 7. We havenot determined whether the heteroduplex pattern ob-served in both sample groups for exon 5 is the same

reported polymorphism. We did not observe any evi-dence of sequence variation in the control group forexon 7. The most-common polymorphism observedwas in exon 3 and causes an amino acid change Serl6-Leu. It was found to have a heterozygosity ap-

proaching 50%. This can be typed by digestion of theexon 3 PCR product with Sau3AI. The 16Leu alleleis refractory to digestion, while the 16Ser allele digeststo 169-bp and 137-bp fragments.

#1062 #1515

Exon 7

Exon 9

Figure 2 EtBr-stained MDETM gel, showing heteroduplexpatterns of exon 7 and 9 mutations. PCR products from families1062 and 1515, which carry the Arg212His and Arg302His muta-

tions. PCR products with (+) and without (-) cosmid 7D11 PCRcontrol DNA were denatured and cooled slowly to allow the for-mation of heteroduplexes prior to gel electrophoresis. Exon 7 PCRproducts were loaded and run for 20 min at 1,200 V. Exon 7products with the addition of 7D11 control were loaded in thesame lanes and the gel was electrophoresed according to Subjectsand Methods. Exon 9 PCR products with and without cosmidcontrol PCR product were electrophoresed in a manner similar to

that for exon 7. Family 1062 shows paternal inheritance of bothArg212His and Arg302His mutations with affected siblings beingheterozygous. Affected sibs in family 1515 are homozygous forboth mutations (note the absence of heteroduplex formation with-out the addition of 7D11 PCR product), with each parent beingobligate heterozygotes.

3' Tail

/.

1078

n bindinMe

2p 072p

0666-i 6 i

Weston et al.: Mutation Analysis of Myosin VIa

#734 #218

B ;

Figure 3 Segregation of exon 13 and exon 14 mutations in EtBr-stained MDETM gels. A, Heteroduplex pattern of exon 13 PCR product(432 bp), showing the 468 + Gln mutation segregating in family 985.The proband, her mother, and maternal grandmother are heterozy-gous for a GCA insertion. The last lane on the right side of panel Ais exon 13 PCR product amplified from 7D1 1. B, Heteroduplex pat-

terns of exon 13 Pro5O3Leu (family 734), Glu450Gln, and the intronicIVS13nt-8c-+g mutations (family 218). The proband from 218 carriestwo different mutations (maternal Glu41OGln, paternal IVS13nt-8c-+g), while the unaffected brother is only heterozygous for IVS13nt-8c-+g mutation. Family 734 is an unusual USH1B family. The sisteris heterozygous for the ProSO3Leu mutation and was diagnosed witha progressive hearing loss (half-filled symbol), probably unrelated to

Usher syndrome.

Discussion

Myosins are a superfamily of molecular motors thatinteract with actin filaments, converting energy fromATP hydrolysis into mechanical force that is used toperform diverse cellular functions (Mooseker andCheney 1995). Myosin VIIa is one of > 10 separateclasses of unconventional myosins that are characterizedas having a conserved N-terminal motor domain, a regu-

latory domain, and a unique tail domain. The recentcharacterization of the myosin VIIa tail reveals a shorta-helical region predicted to dimerize myosin VIIa intoa two-headed molecule (Hasson et al. 1995; Chen et al.1996; Weil et al. 1996), and two repeated units eachcontaining a novel MyTH4 domain similar to domainsin three other myosins and a domain homologous to

talin, a member of the band 4.1 family (Chen et al.1996). Indirect immunofluorescence studies using anti-bodies specific for myosin VIIa protein have determinedits location within tissues important to the pathology ofUsher syndrome. Myosin VIIa was found to be limitedto the inner and outer hair cells of the guinea pig organ

of Corti and to the retinal pigmented epithelium (RPE)in adult rat (Hasson et al. 1995). Using a human-specific

myosin VIIa antibody, El-Amraoui et al. (1996) recentlydemonstrated that myosin VIIa was localized in the hu-man adult retina to the RPE cell layer, as well as therod and cone photoreceptors cells, and confirmed thatexpression in rodents was limited to the RPE. However,it is not clear from these observations that the RP inUSHiB is caused by myosin VIIa defects in the RPE, therods and cones or both cell types. Indeed, the functionof myosin VIIa may be different in the two affectedsensory organs. If so, it is tempting to speculate thattissue-specific differences in function could lead to dif-ferences in the effect of specific mutations in the eyeversus the inner ear. Alternative splicing is one mecha-nism for creating related proteins with varied functions.A number of alternate splice forms of MY07A havebeen found, but the significance of these different formsis not known (Weil et al. 1996; Chen et al. 1996). Alter-natively, myosin VIIa may have a functionally different,although biochemically similar, function in differenttypes of cells. Myosin VIla is highly expressed in Sertolicells of the testis, an obvious nonsensory organ, butthere has been no reported effect of Usher Tb mutationson spermatogenesis or fertility. The hypothesis of func-tional differences of myosin VIIa in the eye versus theinner ear is consistent with the observation that one typeof recessive nonsyndromic hearing loss maps to the sameregion as MY07A on 1 1q. Furthermore, the pattern ofthe disease in the eye versus the ear is different: the RPis partial and progressive with age, but the hearing lossis total and congenital. Consequently, we would predictthat, as the nature and distribution of the MY07A mu-tations are uncovered, the types of mutations and theirdistribution to specific functional domains occurring inthe MY07A gene will be correlated with specific differ-ences in phenotype.While the sample reported on here consists only of

Usher I patients, mutation surveys of MY07A in pa-tients with isolated RP or isolated hearing loss are ex-pected to be carried out. The effect of different muta-tions on the severity of the Usher I phenotype or theircorrelation with atypical cases of Usher I have not yetbeen investigated. While our experience suggests thatthere is little variation in the phenotype of Usher Ib, thisimpression is influenced by a severe bias in the criteriafor the ascertainment of our families for linkage studies.If there were Usher Tb cases with milder hearing losses,normal vestibular responses, or other atypical features,they would not have been included in the original study.We are also in the process of organizing a mutationscreening of atypical cases to compare with the resultsobtained for typical Usher lb cases.

In the absence of a comprehensive biochemical pictureof the role of myosin VIIa in cellular function, it is diffi-cult to predict which mutations are pathologic andwhich are benign. This is especially problematic in a

#985A . s-i

A 5w~

1079

Am. J. Hum. Genet. 59:1074-1083, 1996

A

7VI1

3,

5,

AC a T ACGT ACCo A c of

BMother GI USHlb sibsMC CTG CAG TAT GAG Ggtg MC CTG CAG TAG GAG GgtgGgtgasn lou gin *asn lau gin tyr glu asn lou gin *

7D1 IAAC CTG CAG TAT GAG

asn lou gin tyr glu

Figure 4 Tyr333Stop mutation in exon 9. A, Sequencing gel from family 1325. Exon 9 PCR product from family 1325 was sequencedusing MY07-9L primer (table 1). The sequence is, therefore, from the complementary strand. The 5'- to-3' direction is shown. The arrowpoints to the heterozygous C/A of mother. The corresponding position in the affected siblings are homozygous C. B, Consequences of the basealteration to the amino acid sequence. The vertical black bar to the right of panel A corresponds to the region of the sequence whose complementis shown translated. The G-+T transversion changes a tyrosine to a stop codon.

recessive disorder like Usher Tb, where only a few caseshave been identified, to date, in which both mutant al-leles are known. It is presumed here that all null muta-tions will cause disease in the homozygous state or whenpaired with a different null mutation. Alterations in theamino acid sequence of the myosin VIla polypeptide

Table 3

Polymorphisms Observed in the Myosin Vila PCR Products

Exon-Specific PCR product Polymorphic Change

3 Serl6Leu, TCG-TTG4 IVS3nt-88c-+t4 IVS3nt-7c-t8 IVS7nt-3c-t8 Gly261Gly, GGT-GGC10 IVS9nt-35c-'g10 IVS10nt+75t--c12 IVS1lnt-8t-+c12 IVS12nt+8g-a14 IVS13nt-42t-c14 AsnS35Asn, AAC-AAT

are not always pathologic, and each mutation must beconsidered in light of the predicted disruption of func-tion. This is a problem, since no one is completely sureof the true function of myosin VIIa in sensory cells.However, the availability of the three-dimensional struc-ture of chicken skeletal myosin S1 (Rayment et al.1993b) is an important resource for predicting the func-tional significance of particular domains of any myosingene, since the primary structure of the N-terminal mo-tor portion of the myosin superfamily has been highlyconserved throughout evolution. There was 61% se-quence similarity between myosin VIIa and chicken skel-etal myosin at the amino acid level in the motor domain.A similar strategy for evaluating mutations was de-scribed for 5-cardiac myosin II (MYH7) (Rayment et al.1995), where missense mutations in this myosin genehave been shown to cause autosomal dominant familialhypertrophic cardiomyopathy (FHC). For this disease,the mutated protein appears to exert a dominant nega-tive effect. Thirty-three of 40 MYH7 mutations occurin the motor domain, indicating that alterations in themotor activity are more likely to lead to a viable, hyper-trophic phenotype than tail mutations (Vikstrom andLeinwand 1996). Of the USH1B mutations reported

#1325

1080

Weston et al.: Mutation Analysis of Myosin VIIa

here, only the ATP pocket mutations Arg212His or Cyscorrespond to those domains functionally important incausing FHC.The Arg212His or Cys mutations in exon 7 lie C-

terminal to the ATP binding loop of the myosin VIIamotor domain at the base of the ATP binding pocket.This region is conserved in all myosins from all classes.Mutations in MYH7 have been observed to occur in thesame conserved motif in FHC patients. In particular, theMYH7 mutations, Asn232Ser and Phe244Leu (Ray-ment et al. 1995), lie within the same conserved motifas the Arg212* mutations. Because of the absolute evo-lutionary conservation of residues including codon 212within this region of the ATP binding pocket, it is ex-pected that the Arg212His or the Arg212Cys mutationswould severely impair the function of the motor. Thisregion of the motor is thought to undergo conforma-tional changes during the actin-myosin power stroke,and so improper head movement and/or incomplete con-version of the energy from ATP hydrolysis are possibleconsequences of mutations within this conserved do-main (Watkins et al. 1995).The MYO7A Glu450Gln and +468Gln mutations in

exon 13 are within a highly conserved a-helix domainof the myosin motor involved in actin binding. The+468Gmn mutation occurs in a region of the a-helixadjacent to the domain that contains the Phe513CysFHC mutation. Interpretation based on the three-dimen-sional (3-D) structure of chicken skeletal muscle sug-gested that the MYH7 PheS13Cys mutation may modifythe nature and timing of conformational changes duringthe contractile cycle (Rayment et al. 1995). The positionof MYO7A +468Gln with respect to the MYH7 PheS1-3Cys in the 3-D myosin structure suggests that +468Glnmay cause a similar loss of motor function. The MYO7AGlu450Gln mutation occurs in the molecular cleft thatsplits the 50-kD fragment into the upper and lower do-main. This cleft provides a communication pathway be-tween the nucleotide- and actin-binding sites (Raymentet al. 1995). The ProSO3Leu mutation occurs in a non-conserved region of the motor, although the Pro5O3 isconserved across several, but not all, myosin types(Mooseker and Cheney 1995). By comparison to the 3-D structure of chicken skeletal muscle myosin II, theequivalent residue is ProS29 and produces a prominentbulge within an a-helix that runs from ProS16 toPhe542, and this region has a potential stereospecificinteraction with actin (Rayment et al. 1993a). Specifi-cally, this region occurs on the outer edge of the mole-cule and is implicated in the two-step reaction of bindingof myosin to actin (Taylor 1993). So, while it is possiblethat this is a phenotypically neutral mutation becauseof the lack of primary sequence conservation among theunconventional myosins in this domain, the observedcontacts between actin and myosin in this region imply

that this is a significant alteration. In addition, the origi-nal sh-1 mouse mutation (Arg-Pro) occurs one codonupstream at the equivalent human position 502 (Gibsonet al. 1995). The identification of two pathologic muta-tions in separate species within this region of the motorsuggests that this region is critical for proper function.Furthermore, ProSO3Leu was not observed in any of the192 control chromosomes.The MYO7A Arg3O2His in exon 9 is a problematic

change in Usher Ib patients and illustrates the difficultyin trying to correlate phenotype/genotype associations.The Arg3O2His alteration is in a nonconserved regionof the motor domain. Arginine at codon 302 has beenobserved in MYO7A sequences from human (U17180),mouse (201683A), pig (U34226), and bull frog(U14379) (Gibson et al. 1995; Hasson et al. 1995; Weilet al. 1995). Two of five different MYO7A cDNA clonesisolated from three independent cDNA libraries werefound to have a histidine at position 302 (Chen et al.1996), suggesting that this mutation occurs in normalpopulations. However, all 192 control chromosomeswere negative for the heteroduplex pattern observed in5/378 USH1B chromosomes with Arg3O2His. Neverthe-less, the Arg3O2His change may be neutral with respectto the USH1B phenotype. Alternatively, it may act withother mutations that cumulatively disable the motor.The fact that Arg3O2His was observed independentlytwice in cis with Arg212His suggests this possibility.Alternatively this observation could just as likely be dueto linkage disequilibrium.The IVS13nt-8c--g mutation is analogous to three hu-

man examples of identical -8-bp splice-acceptor muta-tions reported in the literature, (two cases of 3 thalassae-mia, one case of adrenal hyperplasia) (Krawczak et al.1992). Sebillon et al. (1995) analyzed a t-+g mutationat a position 8 bp upstream of the 3' splice site in intron2 of the f-globin gene by an in vitro splicing assay. Invivo, this mutation was observed in trans with aframeshift mutation on the other allele and resulted ina mild -+ thalassemia phenotype, suggesting that nor-mally spliced P-globin mRNA was produced at reducedlevels. In vitro, this mutation decreased the splicing effi-ciency 2-3-fold relative to the wild type and is corre-lated with a reduction of spliceosome formation. Thus,we would predict that the analogous IVS13nt-8c-+g mu-tation would seriously impair the proper constructionof an effective MYO7A mRNA.One case reported here was homozygous for both

Arg212His and Arg3O2His (fig. 2). Because Arg212Hisand Arg3O2His occur in cis, it is plausible that this dou-ble mutation is important to the USH1B phenotype.Double mutations have been reported in cystic fibrosis,with similar difficulties in assessing whether the occur-rence of these mutations in cis are additive or compensa-tory to the phenotype (Savov et al. 1995). Since the

1081

Am. J. Hum. Genet. 59:1074-1083, 1996

entire coding region for MYO7A has yet to be scannedfor additional mutations in these families, it is still too

early to draw any concrete conclusions pertaining to thepossible cooperation of these mutations in cis specific to

the Usher lb phenotype. In general, the USH1B missensemutations will contribute additional information to thegeneral function of the myosin motor by pointing out

critical domains that, when altered, create functionallyinactive molecules.One aspect of the MYO7A mutation search we

wanted to address was the reason for the phenotypicdifferences between sh-i and Usher lb. Three of theseven Sh-i mutations, Arg241Pro (sh-16j), ArgSO2Pro(sh-i), and del646-655 (sb-1 816sB), are known (Gibsonet al. 1995). In the case of sh-i, the defective myosinVII protein causes a phenotypic effect only in the innerear. It appears that similar missense mutations in hu-mans produce both retinal and cochlear degeneration.The two mutations, mouse Arg5O2Pro and human ProS-03Leu, occur in the same domain yet produce differentphenotypic results with respect to the eye. This is consis-tent with the species-specific difference in myosin VIIaprotein localization observed in the retina (El-Amraouiet al. 1996).More quantitative clinical studies of patients and

heterozygotes are needed. There appears to be no overtclinical differences between the cases categorized bythe types of mutations observed (missense, nonsense).Nonsense mutations would produce no protein, andheterozygotes of null alleles have a decreased dosageof protein, which seems sufficient to produce a normalphenotype. A missense mutation, however, could pro-

duce an aberrant protein present in the cells of a het-erozygote. In light of these considerations, one mightexpect that missense heterozygotes would have mildhearing and/or visual problems but that null heterozy-gotes would be asymptomatic. The case in family 734typifies this possibility. Further investigation into theeffects of the gene in the "normal carriers" may behelpful in elucidating the function of MY07A. An ex-

haustive search for MY07A mutations in Usher lbpatients will certainly contribute and augment the vari-ous biochemical and histological studies now under-way to understand the contribution of MY07A to cellfunction.

Three other mutations in MY07A have been reportedelsewhere (Weil et al. 1995). These mutations includetwo nonsense mutations caused by single base substitu-tions carried by two families that were heterozygous,(ArglSOStop and Gln234Stop) occurring in exons 5 and7, respectively. In addition, a 6-base deletion in exon 7,corresponding to the deletions of codons 218-219, were

found as heterozygotes in two separate families, withone family segregating both the ArglSOStop and de1218-219. This report presents 3 new USH1B cases where

mutations have been found in both alleles and raises thetotal number of unique mutations in MYO7A to 15.The most-common Usher Ib mutations observed to

date are Arg212Cys, Arg212His, and Arg3O2His. To-gether, these comprise -50% of the total mutationsobserved, but this is <3% of the total Usher Ib chromo-somes studied. Furthermore, we have not yet observedany linkage disequilibrium between Usher Ib and severaladjacent polymorphic markers used in previous linkagestudies (W. J. Kimberling; unpublished observations).This is consistent with the hypothesis that there are sev-eral independently occurring mutations and no commonUSHiB allele.The length of the MYO7A cDNA is reported to be

7,409 bp or 7,465 bp (Weil et al. 1996; Chen et al.,in press). We screened the first 14 of 49 exons, whichcover 26% of the coding portion of MYO7A. The partof the gene screened includes -3/4 of the N-terminalmotor portion but none of the light chain binding ortail domains. Thirteen unique mutations were detectedby HA in 23 separate USH1B chromosomes. Conse-quently, we expect to find the majority of MYO7Amutations in the tail. If true, this distribution wouldbe different from that observed for MYH7 mutations,where most have been found in the motor head. It ispossible, though less likely, that mutations occur in theintronic regions of the MYO7A gene and/or a commonmutation that results in relatively large deletions of theMYO7A gene exists. In the latter case, many patientswould be hemizygous and no variation would be seenwith heteroduplex analysis.The study of mutations in MYO7A and their rela-

tionship to the Usher I phenotype is just beginning.We expect that further investigations will reveal moreabout the structure and function of the myosin VIImolecule, and continued progress will result in an en-hanced understanding of the molecular pathology ofthis disorder.

AcknowledgmentsWe wish to thank Dr. S. Brown for the use of the ET58

mouse clone. This work was supported by NIH-NIDCD grantP01 DC01813-01 and grant 89-198 from The FoundationFighting Blindness (Hunt Valley, MD). In Holland, the studywas supported by grants from the Dutch Foundation for theHandicapped Child, the Dutch Foundation "De Drie Lichten,"The Heinsius Houbolt Foundation, and De Landelijke Sticht-ing voor Blinden en Slechtzienden. S.G.J. was supported byNIH grant R01 EY05627. M.B.G. was supported by the Penn-sylvania Lions Sight Conservation and Eye Research Founda-tion and the Eye and Ear Foundation of Pittsburgh. H.T. wassupported by NIH grant P01 DK38979, and D.J.0. by NIHNIDCD grant P60 DC00982-06. We want to especially thankall the patients and their families for participating in this study.

1082

Weston et al.: Mutation Analysis of Myosin VIIa 1083

ReferencesChen Z. Hasson T. Kelley PM, Schwender BJ, Schwartz MF,Ramakrishnan M, Kimberling WJ, et al (1996) Molecularcloning and domain structure of human myosin VIla, thegene product defective in Usher syndrome 1B. Genomics 36:440-448

El-Amraoui A, Sahly I, Picaud S, Sahel J, Abitbol M, Petit C(1996) Human Usher lB/mouse shaker-i: the retinal pheno-type discrepancy explained by the presence/absence of myo-sin VIIA in the photoreceptor cells. Hum Mol Genet 5:1171-1178

Gibson F, Walsh J, Mburu P. Varela A, Brown KA, AntonioM, Beisel KW, et al (1995) A type VII myosin encoded bythe mouse deafness gene shaker-1. Nature 374:62-64

Guilford P, Ayadi H, Blanchard S, Chaib H, Le Paslier D,Weissenbach J, Drira M, et al (1994) A human gene respon-sible for neurosensory, non-syndromic recessive deafness isa candidate homologue of the mouse sh-1 gene. Hum MolGenet 3:989-993

Hasson T. Heintzelman MB, Santos-Sacchi J, Corey DP,Mooseker MS (1995) Expression in cochlea and retina ofmyosin VIIa, the gene product defective in Usher syndrometype 1B. Proc Natl Acad Sci USA 92:9515-9819

Kaplan J, Gerber S, Bonneau D, Rozet JM, Delrieu 0, BriardML, Dollfus H, et al (1992) A gene for Usher syndrometype I (USH1A) maps to chromosome 14q. Genomics 14:979-987

Keen J, Lester D, Inglehearn C, Curtis A, Bhattacharya S(1991) Rapid detection of single base mismatches as hetero-duplexes on hydrolink gels. Trends Genet 7:5

Kimberling WJ, Moller CG, Davenport S, Priluck I, BeightonP, Greenburg J, Reardon W, et al (1992) Linkage of Ushersyndrome type I gene (USH1B) to the long arm of chromo-some 11. Genomics 14:988-994

Kimberling WJ, Moller C (1995) Clinical and molecular genet-ics of Usher syndrome. J Am Acad Audiol 6:63-72

Krawczak M, Reiss J, CooperDN (1992) The mutational spec-trum of single base-pair substitutions in mRNA splice junc-tions of human genes: causes and consequences. Hum Genet90:41-54

Mooseker MS, Cheney RE (1995) Unconventional myosins.Annu Rev Cell Dev Biol 11:633-675

Rayment I, Hoden HM, Sellers JR, Fananapazir L, and EpsteinND (1995) Structural interpretation of the mutation in theB-cardiac myosin that have been implicated in familial hy-pertrophic cardiomyopathy. Proc Natl Acad Sci USA 92:3864-3868

Rayment I, Holden HM, Whittaker M, Yohn CB, Lorenz M,Holmes KC, Milligan RA (1993a) Structure of the actin-myosin complex and its implications for muscle contraction.Science 261:58-65

Rayment I, Rypniewski RW, Schmidt-Base K, Smith R, Tom-chick DR, Benning MM, Winkelmann DA, et al (1993b)Three-dimensional structure of myosin subfragment-1: amolecular motor. Science 261:50-58

Savov A, Angelicheva D, Balassopoulou A, Jordanova A,Noussia-Arvanitakis S, Kalaydjieva L (1995) Double mutantalleles: are they rare? Hum Mol Genet 4:1169-1171

Sebillon P. Beldjord C, Kaplan J, Brody E, Marie JA (1995)T to G mutation in the polypyrimidine tract of the secondintron of the human B-globin gene reduces in vitro splicingefficiency: evidence for an increased hnRNP C interaction.Nucleic Acids Res 23:3419-3425

Smith RJH, Lee EC, Kimberling WJ, Daiger SP, Pelias MZ,Keats BJB, Jay M, et al (1992) Localization of two genesfor Usher syndrome type I to chromosome 11. Genomics14:995-1002

Taylor EW (1993) Molecular muscle. Science 261:35-36Vikstrom KL, Leinwand LA (1996) Contractile protein muta-

tions and heart disease. Curr Opin Cell Biol 97-105Watkins H, Seidman JG, Seidman CE (1995) Familial hyper-

trophic cardiomyopathy: a genetic model of cardiac hyper-trophy. Hum Mol Genet 4:1721-1727

Weil D, Blanchard S, Kaplan J, Guilford P., Gibson F, WalshJ, Mburu P, et al (1995) Defective myosin VIla gene respon-sible for Usher syndrome type 1B. Nature 374:60-61

Weil D, Levy G. Sahly I, Levi-Acobas F, Blanchard S, El-Amoroui A, Crozet F, et al (1996) Human myosin VIIAresponsible for the Usher 1B syndrome: a predicted mem-brane-associated motor protein expressed in developing sen-sory epithelia. Proc Natl Acad Sci USA 93:3232-3237